Author Topic: Linux - Build 0.32 - new crash - gdb log (Read 9928 times)

yossarianuk · « **on:** March 16, 2011, 03:05:13 AM »

Hi.

Just tried build 0.32 - i'm now having a new issue (segfaults) to the 0.31 one.

At random I get :-
=-----------

Starting program: /home/morgan/games/paintball2/paintball2
[Thread debugging using libthread_db enabled]
[New Thread 0xb38b2b70 (LWP 2159)]
[Thread 0xb38b2b70 (LWP 2159) exited]
[New Thread 0xb38b2b70 (LWP 2160)]
[New Thread 0xb14d7b70 (LWP 2161)]
[Thread 0xb14d7b70 (LWP 2161) exited]
[New Thread 0xb02ffb70 (LWP 2164)]
[Thread 0xb02ffb70 (LWP 2164) exited]
[New Thread 0xaa76bb70 (LWP 2166)]
[Thread 0xb38b2b70 (LWP 2160) exited]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xaa76bb70 (LWP 2166)]
0xb2e0fde9 in glDeleteTextures () from /usr/lib/nvidia-current/libGL.so.1
-----------------

Is this perhaps a Nvidia driver issue?

I have reinstalled my desktop(s) I'll go back to build 0.30 and see if that also now has issues.

Regards

jitspoe · « **Reply #1 on:** March 16, 2011, 01:51:03 PM »

How often does this happen? Does it happen if you start up a map and do nothing? What about running around shooting walls? The only thing I can think of that would call glDeleteTextures would be dynamic lightmap updates / stainmaps. You could try setting gl_dynamic to 0.

yossarianuk · « **Reply #2 on:** March 27, 2011, 02:52:27 PM »

HI.

Not sure if it occurs if I enter a map and do nothing. Will try soon and report back.

tried setting gl_dynamic to 0 - same issue.

I got this error (without gdb)

0.30 seems o.k (but many servers I can't join due to version..)

Am using latest nvidia in Ubuntu repos (same occurs in Arch with later version)

---------------------------------------
X Error of failed request: BadWindow (invalid Window parameter)
Major opcode of failed request: 2 (X_ChangeWindowAttributes)
Resource id in failed request: 0x3a00002
Serial number of failed request: 13305
Current serial number in output stream: 13306
X Error of failed request: GLXBadContext
Major opcode of failed request: 128 (GLX)
Minor opcode of failed request: 4 (X_GLXDestroyContext)
Serial number of failed request: 13306
Current serial number in output stream: 13306
---------------------------------------

Cheers!

jitspoe · « **Reply #3 on:** March 28, 2011, 08:58:41 AM »

So this is the same crash you were getting with b31?

yossarianuk · « **Reply #4 on:** March 28, 2011, 03:49:27 PM »

Hi - possibly it is actually....

To 100% check its not the nvidia driver I have tried the latest nouveau driver in Arch linux. (its come on now - still no where near as fast as the offical Nvidia driver)

Same crash (slightly different output as its not using the nvidia driver opengl files)

Here's the gdb log when using the nouveau driver.

-------------------------
Starting program: /home/morgan/paintball2/paintball2
[Thread debugging using libthread_db enabled]
[New Thread 0xf77cdb70 (LWP 3321)]
[Thread 0xf77cdb70 (LWP 3321) exited]
[New Thread 0xf65eeb70 (LWP 3337)]
[Thread 0xf65eeb70 (LWP 3337) exited]
[New Thread 0xf0029b70 (LWP 3338)]
[Thread 0xf0029b70 (LWP 3338) exited]
[New Thread 0xef627b70 (LWP 3339)]
[Thread 0xef627b70 (LWP 3339) exited]
[New Thread 0xf49d4b70 (LWP 3341)]

Program received signal SIGSEGV, Segmentation fault.
0xf7bf3d2f in GL_RenderLightmappedPoly () from ./ref_pbgl.so
#0 0xf7bf3d2f in GL_RenderLightmappedPoly () from ./ref_pbgl.so
-------------------------

If I try to backtrace I get
Cannot access memory at address 0xfffebef0

Also with nvidia driver at least just entering a map can trigger it.

yossarianuk · « **Reply #5 on:** March 28, 2011, 03:56:45 PM »

Just got a backtrace after the crash if this helps - log:-

(from nouveau driver)

Starting program: /home/morgan/paintball2/paintball2
[Thread debugging using libthread_db enabled]
[New Thread 0xf77cdb70 (LWP 3434)]
[Thread 0xf77cdb70 (LWP 3434) exited]
[New Thread 0xf66fdb70 (LWP 3436)]
[Thread 0xf66fdb70 (LWP 3436) exited]
[New Thread 0xeff37b70 (LWP 3439)]
[Thread 0xeff37b70 (LWP 3439) exited]
[New Thread 0xf54e8b70 (LWP 3440)]
[Thread 0xf54e8b70 (LWP 3440) exited]
[New Thread 0xef736b70 (LWP 3442)]
[Thread 0xef736b70 (LWP 3442) exited]
[New Thread 0xeef35b70 (LWP 3443)]

Program received signal SIGSEGV, Segmentation fault.
0xf7bf3d2f in GL_RenderLightmappedPoly () from ./ref_pbgl.so
#0 0xf7bf3d2f in GL_RenderLightmappedPoly () from ./ref_pbgl.so
#1 0xf7bf4f29 in R_RecursiveWorldNode () from ./ref_pbgl.so
#2 0xf7bf4e17 in R_RecursiveWorldNode () from ./ref_pbgl.so
#3 0xf7bf4fae in R_RecursiveWorldNode () from ./ref_pbgl.so
#4 0xf7bf4e17 in R_RecursiveWorldNode () from ./ref_pbgl.so
#5 0xf7bf4e17 in R_RecursiveWorldNode () from ./ref_pbgl.so
#6 0xf7bf4e17 in R_RecursiveWorldNode () from ./ref_pbgl.so
#7 0xf7bf4fae in R_RecursiveWorldNode () from ./ref_pbgl.so
#8 0xf7bf4e17 in R_RecursiveWorldNode () from ./ref_pbgl.so
#9 0xf7bf4e17 in R_RecursiveWorldNode () from ./ref_pbgl.so
#10 0xf7bf4fae in R_RecursiveWorldNode () from ./ref_pbgl.so
#11 0xf7bf4fae in R_RecursiveWorldNode () from ./ref_pbgl.so
#12 0xf7bf4e17 in R_RecursiveWorldNode () from ./ref_pbgl.so
#13 0xf7bf5202 in R_DrawWorld () from ./ref_pbgl.so
#14 0xf7beb8f0 in R_RenderView () from ./ref_pbgl.so
#15 0xf7bec4ba in R_RenderFrame () from ./ref_pbgl.so
#16 0x0806c744 in V_RenderView ()
#17 0x0806acfa in SCR_UpdateScreen ()
#18 0x0805fa9b in CL_Frame ()
#19 0x0808f929 in Qcommon_Frame ()
#20 0x080a4191 in main ()

yossarianuk · « **Reply #6 on:** March 29, 2011, 01:38:59 AM »

And just in case the nouveau driver is adding other issues here is a backtrace after a crash with the Nvidia driver:-
--------------------------------------------------
[New Thread 0xaf3ffb70 (LWP 1749)]
[Thread 0xaf3ffb70 (LWP 1749) exited]
[New Thread 0xa6fffb70 (LWP 1750)]
[Thread 0xa6fffb70 (LWP 1750) exited]
[New Thread 0xa89e3b70 (LWP 1752)]
[Thread 0xa89e3b70 (LWP 1752) exited]
[New Thread 0xa65ffb70 (LWP 1754)]
[Thread 0xa65ffb70 (LWP 1754) exited]
[New Thread 0xa55feb70 (LWP 1755)]
[Thread 0xa55feb70 (LWP 1755) exited]
[New Thread 0xa4dfdb70 (LWP 1756)]
[Thread 0xa4dfdb70 (LWP 1756) exited]
[New Thread 0xa45fcb70 (LWP 1757)]
[Thread 0xa45fcb70 (LWP 1757) exited]
[New Thread 0xa3dfbb70 (LWP 1759)]
[Thread 0xa3dfbb70 (LWP 1759) exited]
[New Thread 0xa35fab70 (LWP 1760)]
[Thread 0xb30a9b70 (LWP 1733) exited]

Program received signal SIGSEGV, Segmentation fault.
0xb26514d2 in ?? ()
#0 0xb26514d2 in ?? ()
#1 0xb2651e9c in ?? ()
#2 0xb265ade2 in ?? ()
#3 0xb265c382 in ?? ()
#4 0xb2673076 in ?? ()
#5 0xb26758fa in ?? ()
#6 0xb26764ba in ?? ()
#7 0x0806c744 in V_RenderView ()
#8 0x0806acfa in SCR_UpdateScreen ()
#9 0x0805fa9b in CL_Frame ()
#10 0x0808f929 in Qcommon_Frame ()
#11 0x080a4191 in main ()

--------------------------------------------------

yossarianuk · « **Reply #7 on:** April 05, 2011, 02:09:12 PM »

Build 0.30 100% doesn't have this issue - just played for a prolonged time and there was no crash.

Is there anything else I can do to help diagnose what's wrong with 0.32 ? A fair few servers are not allowing play in lower versions than 0.32 now.

Here is another backtrace (from arch 64bit) nvidia driver - btw in order to get a backtrace I go to a console window (ctrl+atl+f1) and attach gdb , then continue - that seems the only way to get a backtrace.

--------------------------------
[New Thread 0xf4695b70 (LWP 3801)]
[Thread 0xf4695b70 (LWP 3801) exited]
[New Thread 0xef51eb70 (LWP 3802)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xef51eb70 (LWP 3802)]
0xf718f849 in glDeleteTextures () from /usr/lib32/libGL.so.1
#0 0xf718f849 in glDeleteTextures () from /usr/lib32/libGL.so.1
#1 0xf730dc7c in ?? ()
#2 0xf733629e in ?? ()
#3 0x080a3861 in VID_Shutdown ()
#4 0x0805fe05 in CL_Shutdown ()
#5 0x0808d67b in Com_Error ()
#6 0x08090d66 in FS_Read ()
#7 0x08090ebb in FS_LoadFileZ ()
#8 0xf72e3d5b in ?? ()
#9 0xffe5dedc in ?? ()
#10 0xf72e8ce4 in ?? ()
#11 0xffe5dedc in ?? ()
#12 0xf756cdf0 in start_thread () from /usr/lib32/libpthread.so.0
#13 0xf76519be in clone () from /usr/lib32/libc.so.6
A debugging session is active.
--------------------------------

yossarianuk · « **Reply #8 on:** May 09, 2011, 04:35:19 PM »

Hi,

Ever since build 0.31+ I've not been able to play on a Linux client - random crashes - gererally after a min or so (normally when i'm doing well..)

As well as gdb logs in my previous post I did just get this error message (in case this helps)

--------------------
Recursive shutdown.
Error: Recursive error after: FS_Read: 0 bytes read, should have been 4096.
Received signal 11, exiting...
-------------------------

Previous post : http://dplogin.com/forums/index.php?topic=22294.0

Can anyone advise me on providing more info so this bug can be cured ?

i.e - shall I install debug packages in order to get more gdb output? (will that help)

0.30 works perfectly....

p.s - wine works fine but playing in Linux is far more responsive (i'm better at playing too)

I've checked the following

- installed full version + upgraded 0.30 -> 0.32 = same
- Driver - tried the binary nvidia driver and the open source one (nouveau) = same
- Tried full screen / windowed - every single nvidia DVI monitor setting going.
- its not realted to distro - happens on everyone I have tried
- its not a 32/64 bit thing - happens on both arch's (on mutiple distros)

jitspoe · « **Reply #9 on:** May 09, 2011, 05:29:03 PM »

It's kind of interesting that you've tried all these distros and had the issue, and I've tried a couple and not had the issue. Guess that's just par for the course. I think it has something to do with the video drivers. I'll have to try using the same drivers you are and see if I get the issue.

I'm going to merge these threads to keep all the info in one place.

jitspoe · « **Reply #10 on:** June 05, 2011, 10:55:48 PM »

I finally did figure out a way some of the b31+ code could crash, but it involves doing a vid_restart after you've been playing. The call stack is different, though. Unless you're somehow changing gl contexts or drivers or something while the game is running (only way I know of is with a vid_restart), you shouldn't get the crash I found.

I'm still a bit perplexed by this. I just updated the Ubuntu distro on my gaming laptop to 10.01, I think, and it says the nvidia drivers are "current". I thought maybe it was an issue with newer distros, but I'm still not able to repro this crash with normal play. I have an 8800M in one laptop and an integrated intel 810 or something in my other laptop - neither of them have this issue.

I'm pretty sure it boils down to the gl context or function pointers getting screwed up, but I don't know why. I've done some rearrangements to the rendering code to make it more manageable across OS's, but, functionally, it should be the same. There was a definite issue in b31, but that should have been fixed for b32. I guess I'll keep digging.

yossarianuk · « **Reply #11 on:** June 07, 2011, 04:54:23 PM »

Hi.

Thanks for your investigations jitspoe.

Still having to play with 0.30 for the time being - the crash in 0.32 occurs frequently never in 0.30

Just to give you more info on my setup:-

Usually I am using 1920 x 1080 as my desktop res and 1440x900 for game (the dvi setting is set to display at original native resolution - i.e un stretched - however crash occurs on whatever
setting in when using version higher than 0.30)

Let me know if I can do anything to help.

I am going to be upgrading my nvidia card son (to 400 series) - I'll let you know if it works after I get it In case its some weird anomaly with my present card (nvidia 8500)

Many thanks

jitspoe · « **Reply #12 on:** June 08, 2011, 10:30:46 AM »

Would it be possible for you to try a vanilla Ubuntu 10 configuration with the vanilla Paintball2 configuration to see if you still get the crash?

yossarianuk · « **Reply #13 on:** June 09, 2011, 02:05:11 PM »

Hi

Do you mean Ubuntu 10.04 or 10.10 (or 11.04) ?

Also When you say vanilla am I o.k to install the Nvidia driver (i.e hardware drivers in Ubuntu) ?

thanks

jitspoe · « **Reply #14 on:** June 10, 2011, 03:46:57 PM »

I guess 10.04 - looks like that's the 10.x version they have on their site, and, yeah, install the hardware drivers that are made available through Ubuntu.

yossarianuk · « **Reply #15 on:** June 18, 2011, 10:39:05 AM »

Hi

Unfortunately going back to 10.04 did not help.

I had the amd64 version so used that (crash happens on present version regardless of arch)

Just to confirm - all I did was install 10.04 - install updates - install deps - play painball2

- in paintball I changed resolution to 1440x900 and put it in full screen...

I had not used nvidia-settings at all...

Here are some errors / gdb logs from a crash:-

------------
Received signal 11, exiting...
X Error of failed request: BadValue (integer parameter out of range for operation)
Major opcode of failed request: 130 (XFree86-VidModeExtension)
Minor opcode of failed request: 10 (XF86VidModeSwitchToMode)
Value in failed request: 0x15a
Serial number of failed request: 127377
Current serial number in output stream: 127380
Received signal 11, exiting...
Segmentation fault
------------

Continuing.
[New Thread 0xee90ab70 (LWP 2147)]
[Thread 0xee90ab70 (LWP 2147) exited]
[New Thread 0xedd07b70 (LWP 2149)]
[Thread 0xedd07b70 (LWP 2149) exited]
[New Thread 0xed506b70 (LWP 2150)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xed506b70 (LWP 2150)]
0xf714f969 in glDeleteTextures () from /usr/lib32/nvidia-current/libGL.so.1

And here is another crash in case that helps:-
-------------------------------
Received signal 11, exiting...
X Error of failed request: GLXBadContext
Major opcode of failed request: 128 (GLX)
Minor opcode of failed request: 4 (X_GLXDestroyContext)
Serial number of failed request: 63095
Current serial number in output stream: 63095
X Error of failed request: BadWindow (invalid Window parameter)
Major opcode of failed request: 2 (X_ChangeWindowAttributes)
Resource id in failed request: 0x3400002
Serial number of failed request: 63094
Current serial number in output stream: 63095
-------------------------------

yossarianuk · « **Reply #16 on:** July 07, 2011, 12:22:41 PM »

I have tried switching from DVI to a VGA cable (just in case that made a difference)

I didn't ....

------------------------
Received signal 11, exiting...
X Error of failed request: BadValue (integer parameter out of range for operation)
Major opcode of failed request: 129 (XFree86-VidModeExtension)
Minor opcode of failed request: 10 (XF86VidModeSwitchToMode)
Value in failed request: 0x15a
Serial number of failed request: 53882
Current serial number in output stream: 53885
Received signal 11, exiting...
Segmentation fault
-----------------------

GDB:-

-----------------
ontinuing.
[New Thread 0xf42a6b70 (LWP 1431)]
[Thread 0xf42a6b70 (LWP 1431) exited]
[New Thread 0xeea75b70 (LWP 1432)]
[Thread 0xeea75b70 (LWP 1432) exited]
[New Thread 0xee274b70 (LWP 1434)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xee274b70 (LWP 1434)]
0xf71c13a9 in glDeleteTextures () from /usr/lib32/libGL.so.1
#0 0xf71c13a9 in glDeleteTextures () from /usr/lib32/libGL.so.1
#1 0xf7347c7c in ?? ()
#2 0xf737029e in ?? ()
#3 0x080a3861 in VID_Shutdown ()
#4 0x0805fe05 in CL_Shutdown ()
#5 0x0808d67b in Com_Error ()
#6 0x08090d66 in FS_Read ()
#7 0x08090ebb in FS_LoadFileZ ()
#8 0xf731dd5b in ?? ()
#9 0xffd12f0c in ?? ()
#10 0xf7322ce4 in ?? ()
#11 0xffd12f0c in ?? ()
#12 0xf75a7c77 in start_thread () from /usr/lib32/libpthread.so.0
#13 0xf769150e in clone () from /usr/lib32/libc.so.6
A debugging session is active.

-----------------

ViciouZ · « **Reply #17 on:** July 07, 2011, 01:05:34 PM »

Do some of these happen while the game is loading?

yossarianuk · « **Reply #18 on:** July 07, 2011, 02:10:12 PM »

Hi

No - Its randomally whilst playing - usually within a few minutes.

Exact same settings fine with build 0.30 - the crash never occurs.

jitspoe · « **Reply #19 on:** July 26, 2011, 06:22:06 PM »

I'm trying to repro this on another computer, but I'm having difficulty getting the 32bit x11 libs.

http://dplogin.com/forums/index.php?topic=23410.0